772 Appendix L of compression. Such an engine uses the same substance over and over again, and the cycle is strictly reversible, for we have complete expansion to <r, and complete compression to a. Its u h: k lV,7 i efficiency therefore is measured like the cycle of Carnot. substance is water at a, and steam and water at other times. he P. 623. Ratio of Expansion in a Single Cylinder.— As a result of many recent practical experiments, it is found that a ratio of expansion of between 6 and 9 is the highest limit which can be adopted with economy, though a ratio of 4 will yield practically as good results. P. 624. Theoretical Diagram.-—In order to make the cal- culations from the preliminary diagram agree with practical diagrams, Prof. Unwin introduces a fractional coefficient in the formulge, which he calls the diagram factor, and which is deduced from experimental results of engines similar to that under con- sideration. The value of the factor varies from '85 to '95 in good engines. P. 637. Exhaust Lap.—In the case of a D valve, an early cut-off to steam will catfse an early cut-off to exhaust (see Zeuner's diagram, p. 660). In such a case, if a later compression point be advisable, it may be necessary njpt only to eliminate exhaust lap entirely, but actually to give a small opening to exhaust when the valve is at mid stroke. Such opening is then termed negative lap, and would be shewn on the steam circle in Fig. 653.